snapshot. prepare to debug

diff --git a/src/caffe/optimization/solver.cpp b/src/caffe/optimization/solver.cpp
index d9ab2c1..fb38237 100644 (file)
--- a/src/caffe/optimization/solver.cpp
+++ b/src/caffe/optimization/solver.cpp
@@ -114,6 +114,9 @@ void SGDSolver<Dtype>::ComputeUpdateValue() {
   vector<shared_ptr<Blob<Dtype> > >& net_params = this->net_->params();
   // get the learning rate
   Dtype rate = GetLearningRate();
+  if (this->param_.display() && this->iter_ % this->param_.display() == 0) {
+    LOG(ERROR) << "Iteration " << this->iter_ << ", lr = " << rate;
+  }
   Dtype momentum = this->param_.momentum();
   Dtype weight_decay = this->param_.weight_decay();
   // LOG(ERROR) << "rate:" << rate << " momentum:" << momentum

diff --git a/src/programs/imagenet.prototxt b/src/programs/imagenet.prototxt
index 65b7432..d5e9b92 100644 (file)
--- a/src/programs/imagenet.prototxt
+++ b/src/programs/imagenet.prototxt
@@ -7,7 +7,7 @@ layers {
     batchsize: 96
     subtraction: 114
     cropsize: 227
-    mirror: true
+    mirror: false
   }
   top: "data"
   top: "label"

diff --git a/src/programs/imagenet_solver.prototxt b/src/programs/imagenet_solver.prototxt
index 58b0dfe..6e58363 100644 (file)
--- a/src/programs/imagenet_solver.prototxt
+++ b/src/programs/imagenet_solver.prototxt
@@ -1,10 +1,10 @@
-base_lr: 0.02
+base_lr: 0.01
 lr_policy: "step"
 gamma: 0.1
 stepsize: 340000
-display: 100
+display: 20
 max_iter: 1200000
 momentum: 0.9
 weight_decay: 0.0005
-snapshot: 15000
+snapshot: 5000
 snapshot_prefix: "alexnet_train"
\ No newline at end of file